Why Reconciliation Comes Before the Expense Calculation
An expense-per-exposure calculation is only as sound as the two quantities it combines. The numerator must represent the expense applicable to the business being priced, and the denominator must represent the corresponding exposures over the same period. A ledger total can be accurate while the resulting provision is wrong because the policy statistics omit months, include duplicate records, or cover a different set of policies.
This tutorial focuses on those data checks. It does not change the expense method: it establishes whether expense dollars and policy statistics are comparable enough to use in one. The earlier material on policy counts and exposure allocation distinguishes a policy count from an exposure measure. For example, 100 policies in force for one month each represent 100 policy-months, not 100 policy-years.
A policy count is a count of policies at a specified point in time or during a specified interval. An exposure measure accumulates the time or units of insurance represented by those policies. If the selected unit is a policy-year, 12 policy-months equal one policy-year. This conversion does not turn a point-in-time policy count into an exposure measure: the duration represented by the statistics must still be known.
The word matching carries most of the analytical burden. A full-year expense total paired with nine months of exposure will generally overstate expense per exposure. Expenses for several coverages divided by the exposures for only one coverage will also generally overstate that coverage’s provision. Conversely, extra or duplicated exposures can understate the calculated amount. Reconciliation is not simply a clerical tie-out: it tests whether the comparison answers the intended pricing question.
A Practical Reconciliation Sequence
Before calculating a provision, state the intended unit of analysis: for example, dollars of a specified expense category per policy-year for a specified coverage and experience period. Then test the source data against that definition. The following sequence makes the checks explicit.
Name the expense category, coverage or business segment, period, and exposure unit required for the calculation.
Check the ledger dates, accounting adjustments, expense categories, and whether the total includes business outside the target population.
Check that the policy statistics cover the same interval and that counts, policy-months, and policy-years are not being confused.
Check for omissions, duplicates, cancellations, late transactions, inconsistent identifiers, or policy records assigned to the wrong segment.
Make only supported adjustments, retain a traceable reconciliation, and divide the aligned expense total by the aligned exposure measure.
The reconciliation should preserve the distinction between a difference that has been explained and one that has merely been forced to balance. For example, an expense ledger may include a late invoice for work performed in the target period. Whether that invoice belongs in the selected expense total depends on the expense measurement basis and period definition; its posting date alone may not settle the question. Likewise, a policy statistic may change because of a corrected effective date, not because an exposure should be added without review.
Worked Applications
Worked Example: Completing a Full-Year Exposure Denominator
A hypothetical insurer has $180,000 of applicable expenses for a full calendar year. Its policy-statistics extract contains 720 policy-months for the first nine months. A validated supplemental extract adds 240 policy-months for the remaining three months. Calculate the full-year expense per policy-year, and show the result if the incomplete denominator is mistakenly used.
Method and alignment check. The expense total covers 12 months, so the exposure denominator must cover the same 12 months. The supplemental extract is validated and covers the missing final three months.
First add the policy-months:
Convert the full-year exposure to policy-years:
Now divide the full-year expense total by those 80 policy-years:
For comparison, using only the first nine months gives \(720/12=60\) policy-years and \(\$180{,}000/60=\$3{,}000.00\) per policy-year. That calculation pairs a full-year numerator with an incomplete denominator; it is not the aligned full-year result. The incomplete-denominator result is $750.00 higher per policy-year, or one-third higher than $2,250.00.
Check. The full-year denominator exceeds the nine-month denominator, while the expense numerator is unchanged. Therefore its expense per policy-year must be lower: $2,250.00 is less than $3,000.00. The common mistake is to treat 720 policy-months as though they represented the entire year, or to divide dollars by a raw policy-month count without converting to the requested unit.
Worked Example: Matching Expense Scope to Coverage Exposures
A hypothetical insurer records $150,000 of policy-servicing expenses for two coverage groups combined. A documented activity-based allocation assigns 60% of the total to Coverage A and 40% to Coverage B. The matching exposure statistics are 600 policy-years for Coverage A and 400 policy-years for Coverage B. Calculate the expense per policy-year for each coverage and compare it with the result of dividing all expenses by Coverage A exposures.
Method and alignment check. The ledger total covers both groups, but the requested provisions are coverage-specific. Apply the documented expense allocation before dividing by each group’s exposures.
Allocate the expense total:
The allocated amounts reconcile to the ledger total: \(\$90{,}000+\$60{,}000=\$150{,}000\). Divide each allocated amount by its matching policy-years:
If the full $150,000 were divided by Coverage A’s 600 policy-years, the result would be \(\$150{,}000/600=\$250.00\) per policy-year. That amount assigns Coverage B’s $60,000 of expenses to Coverage A. It is not a valid Coverage A provision under the stated allocation.
Check. The two coverage results are equal because the allocated expense amounts are $150 per policy-year in each group. Together, the exposure-weighted expenses are \(600\times \$150+400\times \$150=\$150{,}000\), which matches the ledger total. The common mistake is to assume that a total expense amount is automatically in scope for whichever exposure denominator is available.
Worked Example: Correcting Duplicate and Missing Policy-Months
A hypothetical insurer’s annual ledger shows $200,000 of expenses for one business segment. The policy-statistics extract reports 14,100 policy-months. An audit finds that the extract includes 300 duplicated policy-months and omits 1,200 policy-months for valid December records. The duplicate and omitted records are distinct, and all the identified records belong to the target segment and year. Calculate the corrected expense per policy-year and the uncorrected result.
Method and alignment check. Correct the exposure count before converting it to policy-years. Subtract the duplicate records and add the omitted valid records:
Convert the corrected policy-months to policy-years:
The corrected expense per policy-year is:
If neither issue were corrected, the reported 14,100 policy-months would equal \(14{,}100/12=1{,}175\) policy-years. The uncorrected calculation would be \(\$200{,}000/1{,}175=\$170.212765\ldots\), or $170.21 per policy-year rounded to the nearest cent. This is $10.21 higher than the corrected result.
Check. The net correction adds \(1{,}200-300=900\) policy-months, or \(900/12=75\) policy-years. A larger denominator with the same expense numerator must produce a lower expense per policy-year, as it does. The common mistake is to notice the missing December records but overlook duplicates, or to adjust one count without checking whether the records overlap.
Why the Matching Basis Matters
The expense-per-exposure formula follows from the meaning of a unit rate. If \(E\) is the applicable expense total and \(M\) is the matching number of policy-months, then the number of policy-years is \(M/12\). Dividing expense by that number gives the amount associated with one policy-year:
This equality is an algebraic conversion, but its interpretation depends on alignment. If the expense total covers a full year and \(M\) includes only nine months, then \(E\) and \(M/12\) do not describe the same exposure period. The arithmetic still returns a number, but that number is not the full-year expense per policy-year requested. Similarly, if \(E\) includes two coverages while \(M\) contains only one, the result attributes expenses outside the denominator’s scope to the included coverage.
If both numerator and denominator are consistently restricted to the same subset or interval, their ratio can still be meaningful for that restricted basis. For example, an eight-month expense total divided by eight-month policy exposure can produce an expense per policy-year equivalent. But it should not be described as a full-year observation unless the data have been adjusted and the adjustment is supported. A ratio alone cannot repair an inconsistent period or scope.
| Potential mismatch | Diagnostic question | Typical consequence if unresolved |
|---|---|---|
| Different periods | Do expenses and exposures cover the same dates, including late adjustments? | A partial exposure denominator may make the expense rate too high; the direction depends on the specific mismatch. |
| Different business scope | Do expenses and exposures refer to the same coverage, segment, or policy population? | Costs may be attributed to exposures that did not generate or receive them. |
| Different units | Are statistics policy counts, policy-months, or policy-years? | The calculated provision may be off by a duration conversion or otherwise have the wrong unit. |
| Incomplete or duplicated records | Are valid records missing, duplicated, or assigned to the wrong period? | A denominator that is too small generally increases the calculated rate; one that is too large generally decreases it. |
The direction in the final row assumes the numerator is fixed and positive. More generally, if \(E>0\), then \(E/D\) decreases as denominator \(D\) increases. That observation is a useful reasonableness check, not a substitute for determining which records belong in the denominator.
Common Reconciliation Traps
- Using an available denominator without testing its scope. A convenient policy count may exclude a coverage or include a segment that is not represented by the expense total.
- Confusing policy counts with exposure units. A count of policies at year-end does not necessarily equal the number of policy-years accumulated during the year.
- Pairing a full-period expense with partial-period statistics. As the first example shows, this can materially overstate expense per exposure.
- Forcing totals to agree without evidence. A balancing adjustment should be supported by a source record or documented allocation, not added merely to eliminate a difference.
- Correcting only one kind of record error. Missing records and duplicated records can coexist and offset partly or fully; investigate both.
Check Your Understanding
For each problem, reconcile the stated information before calculating the requested amount.
- A hypothetical insurer has $72,000 of annual expenses and 4,800 matching policy-months. What is the expense per policy-year? (Answer: $180.00 per policy-year)
- A full-year expense total is $126,000. The policy statistics contain 420 policy-years through the end of September, with 140 additional policy-years validated for the final quarter. What is the aligned full-year expense per policy-year, and what would the incomplete denominator produce? (Answer: $225.00 using 560 policy-years; $300.00 using 420 policy-years)
- A $240,000 expense total covers two hypothetical coverage groups. A documented allocation assigns 35% to Coverage X, which has 700 policy-years. What is Coverage X’s expense per policy-year? (Answer: $120.00 per policy-year)
- An extract reports 9,600 policy-months, including 600 duplicated policy-months, and omits 1,200 valid policy-months. The matching annual expense total is $153,000. What is the corrected expense per policy-year? (Answer: $180.00 per policy-year)
- A hypothetical insurer records $96,000 of applicable expenses for an eight-month period. Matching statistics for the same period and business scope total 5,600 policy-months. What is the expense per policy-year equivalent? (Answer: $205.71 per policy-year, rounded to the nearest cent)